Scholarlink

Background <p>In patients with anterior circulation large vessel occlusion stroke, hyperdense areas (HDAs) on post-thrombectomy computed tomography (CT) reflect blood–brain barrier injury and may herald malignant edema or hemorrhagic transformation. Both complications can progress to life-threatening mass effect, often necessitating surgical interventions such as decompressive craniectomy. We aimed to develop and validate a location-specific HDA score for predicting severe mass effect.</p> Methods <p>This multicenter retrospective study analyzed 865 patients exhibiting HDAs on post-thrombectomy CT. The primary outcome was severe mass effect, defined as decompressive craniectomy or midline shift ≥ 5&#xa0;mm on follow-up imaging. A location-specific HDA score was developed by expanding the ASPECTS template to include subarachnoid and ventricular compartments. Predictive models integrating imaging and clinical variables were constructed via LASSO regression and validated internally (bootstrapping) and externally.</p> Results <p>Three HDA locations independently predicted the outcome: insula [adjusted odds ratio (aOR) = 2.02, 95% confidence interval (CI) 1.07–3.73, <i>P</i> = 0.030], subarachnoid space (aOR = 2.47, 95% CI 1.30–4.91, <i>P</i> = 0.005), and ventricles (aOR = 4.58, 95% CI 1.33–15.80, <i>P</i> = 0.016). The HDA score demonstrated a strong dose–response relationship with outcome risk, with scores ≤ 3 indicating low risk (&lt; 10% probability) of severe mass effect and a high negative predictive value (88.8%). An integrated model (model 2) incorporating HDA features, admission blood glucose, National Institutes of Health Stroke Scale (NIHSS) and ASPECTS significantly outperformed a baseline clinical model (model 1) [area under the curve (AUC) 0.801 vs. 0.737, <i>P</i> &lt; 0.001], showing high specificity (81.3%). Model performance was maintained in external validation (AUC 0.786) and in the contrast-staining subgroup (AUC 0.808).</p> Conclusions <p>A simple, location-specific HDA score, especially when integrated into a clinical prediction model, effectively identifies patients at high risk for severe mass effect after thrombectomy. This tool enables early risk stratification and may guide timely intensive care management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Location-Specific Hyperdense Area Score Predicts Severe Mass Effect after Thrombectomy

  • 期刊论文

  • 发表日期：

    2026年5月23日

查看全文
  • SiShan Wen,
  • Hao Tang,
  • JiaNan Li,
  • QiJi Jin,
  • Xi Shen,
  • Lulu Xiao,
  • AnYu Liao,
  • WuSheng Zhu,
  • ChangSheng Zhou,
  • Liang Pan,
  • Jie Chen,
  • Wei Xing,
  • Qian Chen,
  • XinDao Yin,
  • Duchang Zhai,
  • JianRui Li,
  • Wu Cai,
  • XiaoQing Cheng

摘要

Background

In patients with anterior circulation large vessel occlusion stroke, hyperdense areas (HDAs) on post-thrombectomy computed tomography (CT) reflect blood–brain barrier injury and may herald malignant edema or hemorrhagic transformation. Both complications can progress to life-threatening mass effect, often necessitating surgical interventions such as decompressive craniectomy. We aimed to develop and validate a location-specific HDA score for predicting severe mass effect.

Methods

This multicenter retrospective study analyzed 865 patients exhibiting HDAs on post-thrombectomy CT. The primary outcome was severe mass effect, defined as decompressive craniectomy or midline shift ≥ 5 mm on follow-up imaging. A location-specific HDA score was developed by expanding the ASPECTS template to include subarachnoid and ventricular compartments. Predictive models integrating imaging and clinical variables were constructed via LASSO regression and validated internally (bootstrapping) and externally.

Results

Three HDA locations independently predicted the outcome: insula [adjusted odds ratio (aOR) = 2.02, 95% confidence interval (CI) 1.07–3.73, P = 0.030], subarachnoid space (aOR = 2.47, 95% CI 1.30–4.91, P = 0.005), and ventricles (aOR = 4.58, 95% CI 1.33–15.80, P = 0.016). The HDA score demonstrated a strong dose–response relationship with outcome risk, with scores ≤ 3 indicating low risk (< 10% probability) of severe mass effect and a high negative predictive value (88.8%). An integrated model (model 2) incorporating HDA features, admission blood glucose, National Institutes of Health Stroke Scale (NIHSS) and ASPECTS significantly outperformed a baseline clinical model (model 1) [area under the curve (AUC) 0.801 vs. 0.737, P < 0.001], showing high specificity (81.3%). Model performance was maintained in external validation (AUC 0.786) and in the contrast-staining subgroup (AUC 0.808).

Conclusions

A simple, location-specific HDA score, especially when integrated into a clinical prediction model, effectively identifies patients at high risk for severe mass effect after thrombectomy. This tool enables early risk stratification and may guide timely intensive care management.